See how TSO Logic
simplifies IT transformation.

Our software platforms remove the guesswork so you can now plan with confidence.

reCAPTCHA is required.

See how TSO Logic
simplifies IT transformation.

Our software platforms remove the guesswork so you can now plan with confidence.

reCAPTCHA is required.

The Economics Behind Cloud Migration and Enterprise Transformation

Is it more economical to run workloads on premise or in the cloud? Should you invest in refreshing technology or cloud migration? With diverse workloads running across thousands of servers, it can be hard to find empirical answers. Typically, businesses rely on static spreadsheets and high-level financial models—at best, an educated guess.

What kinds of insights can be revealed by applying fine-grained, algorithmic analysis to the actual data? TSO Logic decided to find out.

TSO Logic performed a statistical analysis of operating system (OS) instances across a sampling of its North American customer base, representing multiple industries. Using anonymized data, that included hundreds of millions of data points, from over 10,000 physical servers (includes both hypervisors and non-hypervisors) and 25,000 virtual OS instances, we compiled a representative sample to answer a series of questions:

  • What is the cost of running OS instances as currently provisioned?
  • How much is each instance used historically?
  • How well do currently provisioned resources match up to utilization?
  • And, which of these workloads would be economically more viable in future-state environments including cloud or server refresh?

Economics of the Cloud
Available data from the pool of 25,000 OS instances(1) currently running on-premise reveals that:

  • 45 percent of virtualized OS instances could run more economically in the cloud(2)—migrated as-is, without making any other changes to cloud-optimize those applications.
  • Migrating those instances would generate an overall cost savings of 43 percent annually.

Based on historical utilization patterns for the thousands of workloads sampled:

  • 26 percent of current OS instances are over-provisioned based on historical utilization.
  • Right-sizing those instances—migrating them to cloud instances sized for actual utilization—would generate a 36 percent cost reduction compared to provisioning them in the cloud exactly as they’re provisioned on-premises.

Economics of Server Refresh
According to an anonymized statistical analysis of 10,000 physical servers(3) across the TSO Logic North American customer base:

  • 25 percent of servers were more than three years old.
  • Accounting solely for the processing power gains of newer hardware, the same workloads running today on Generation-5 servers could run on 30 percent fewer Generation-9 servers.

That server consolidation is based on provisioning “like-for-like”—deploying hardware to support the same resources per OS instance that are provisioned right now.

  • By updating older servers to modern hardware and provisioning them based on historical utilization, organizations could support the same workloads with 54 percent fewer servers.

Inside the Data
Any organization considering a cloud migration, or even a hardware refresh, asks the same question: How much will I actually save by doing this? Often though, the answers are little more than an educated guess or back-of-the-napkin calculation. It’s not hard to compare price points for cloud resources for a handful of workloads. But how do you analyze thousands or tens of thousands?

The TSO Logic platform creates a fine-grained statistical model of compute resources and historical utilization patterns to determine the most cost-effective place to run each workload. Ingesting millions of data points, it algorithmically profiles compute patterns. It associates every compute instance with the the physical system on which it’s provisioned and the historical usage and utilization of those resources, down to the processor and provisioned configuration. It then uses pattern matching to determine the best fit for each workload from thousands of potential options. Using up-to-date information from Intel and cloud providers, it normalizes and equates the processing capabilities between various generations of Intel processors and the myriad of cloud options.

Data-Driven Decisions
These numbers are only a top-line analysis. They don’t take into account the other advantages of cloud migration, or changes that organizations can make to applications to make them even more economical They do, however, demonstrate that there is likely significant low-hanging fruit that you could take advantage of to support your application footprint more cost-effectively. If you can identify and migrate those instances, you stand to realize significant financial gain.

Cloud migration, and even hardware refreshes, should not be viewed as an all-or-nothing game. But by better understanding how you’re using current resources, you can make smarter decisions about your cloud strategy and data center investments.

Learn More
To find out how efficiently your organization is using its compute resources, contact us here or email Andrea at

(1) Included in cost calculations for on-premise costs include server amortization, licensing, maintenance and power. Labor is not included in this analysis. Virtual instances reported on come out to an average guest: host ratio of 10:1. (2) Cloud catalog comparison and pricing data is based on Amazon Web Service (AWS) and Microsoft Azure. AWS T-Series instances represent a large portion of the cloud savings. (3) On-premise OS instance cost is based on the amount of resources provisioned to a virtual machine, licensing, maintenance, hardware amortization and power. Physical servers reported contain both hypervisors and logical servers.